Elevated NGF provokes decidual lipid peroxidation and promotes preterm birth in mice

Abstract

Background

Preterm birth (PTB) is a major cause of neonatal morbidity and mortality worldwide, with infection and inflammation being the most common triggers. However, many cases of preterm labor have unknown causes. The maternal decidua is a highly dynamic and heterogeneous region, serving as the nexus at the maternal-fetal interface, connecting the mother and the fetus. Increasing evidence suggests that the maternal decidua plays a crucial role in the initiation of labor. Nerve growth factor (NGF), is an important member of the neurotrophin family and is identified to play a crucial role in initiating the decidual response.

Methods

To investigate whether NGF contributes to preterm birth via lipid peroxidation-dependent pathways, we selected both NGF and erastin (a pharmacological lipid peroxidation inducer) for parallel experimental treatments to examine how these pathways mediate the initiation of parturition. Mice were administered intraperitoneal injections of NGF and erastin. Gestational durations less than 19.5 days were classified as preterm birth. This study employed biological technologies and experimental methods to explore the initiation of delivery and the associated signaling pathways.

Results

Elevated NGF levels in late-stage pregnancy increased the incidence of preterm birth in mice, independent of decidual senescence, placenta abnormal structure and ovarian dysfunction. Instead, NGF treatment activated lipid peroxidation and upregulated inflammatory markers in maternal decidua, particularly cyclooxygenase enzymes, which are critical for labor initiation. Notably, administration of erastin corroborated these findings, leading to similar outcomes in preterm labor.

Conclusions

This study reveals the pivotal role of NGF signaling in promoting excessive lipid peroxidation to disrupt decidual homeostasis and ultimately triggering preterm labor in mice.

Graphical Abstract

Introduction

Preterm birth remains a critical global health challenge. It is a leading cause of neonatal mortality and the second leading cause of death in children under 5 years of age [1, 2]. Beyond infancy, preterm birth contributes to higher morbidity and mortality rates among newborns, with long-term effects including an increased risk of childhood lung diseases and neurodevelopmental disorders, leading to substantial healthcare costs [3, 4]. While several risk factors such as infection, stress, decidual senescence have been identified, the underlying causes of the majority of preterm birth remain unexplained [5, 6]. Inflammatory responses, arising from infections, autoimmune diseases, or other underlying factors, play a significant role in the incidence of preterm birth [7, 8].

Nerve Growth Factor (NGF) is a neurotrophic factor that supports the growth, survival, and differentiation of neurons, while also playing a significant role in immune responses and inflammation [9,10,11,12,13]. NGF manifests in two forms: the mature form, produced through proteolytic cleavage of the precursor form (proNGF), which harbors biological activity and imparts pro-apoptotic and neurotrophic effects [14]. Both NGF forms exert their biological influence by binding to the receptor tyrosine kinase A (TrkA), a prototypical tyrosine kinase receptor [15], as well as the low-affinity p75 neurotrophin receptor (p75NTR) [16]. Although initially investigated in neuronal cells, NGF is acknowledged for its significant roles in other types of cells from various organizations, such as the endocrine, immune, and reproductive systems [17].

The microenvironment at the maternal-fetal interface is crucial for the initiation and maintenance of pregnancy [18], based on the role of NGF in inflammation, a comprehensive understanding of its spatiotemporal dynamics at this interface becomes essential for deciphering pregnancy establishment mechanisms. Previous studies demonstrate that abnormal NGF levels (either elevated or suppressed) disrupt pregnancy homeostasis, resulting in inflammatory responses and increased risks of fetal resorption or spontaneous abortion [19]. This pathological sensitivity correlates with NGF/TrKA expression patterns observed in gestational tissues: during early pregnancy, their predominant expression in the endometrium and fetal membranes drives vascular proliferation, whereas in advanced stages (as evidenced in camelids), this system transitions to vascular remodeling [20]. Notably, peak NGF expression occurs specifically in the decidua during early implantation [21], a compartmentalized pattern further validated in the CBA/J × DBA/2 abortion-prone murine model. In this model, significant upregulation of NGF and TrkA is localized to the decidua rather than the placenta [22], reinforcing the spatial precision of their biological functions. In addition, NGF binding to the receptors on inflammatory cells triggers the release of inflammatory mediators such as histamine, chemokine, cyclooxygenase2 (COX2), prostaglandin D2 and cytokine [16]. This pro-inflammatory effect could represent an additional mechanism through which NGF treatment influences unfavorable pregnancy outcomes. Similarly, the administration of NGF-neutralizing antibodies in normal pregnant mice precipitated abortions and was correlated with an augmented infiltration of NK cells expressing TrkA receptor into the decidua [19]. These observations suggest that maintaining a specific threshold of NGF expression is vital for early pregnancy progression.

NGF does not change at the systemic level during normal pregnancy [23], previous studies on NGF in adverse pregnancy primarily focused on the period around embryo implantation. However, the mechanisms by which high level of NGF production, leads to adverse pregnancy outcomes remain insufficiently understood. In our study, by administering exogenous NGF to pregnant mice in late pregnancy, we found that excessive NGF can induce preterm birth. Notably, these preterm mice did not exhibit abnormalities in placental structure or changes in ovarian function, and the decidual senescence was found to be comparable between NGF-pretreated and control mice. Interestingly, the inflammatory response at the maternal-fetal interface was elevated. Consecutively, we identified that increased lipid peroxidation levels in the decidua, associated with NGF, were linked to preterm birth. Our findings in the present study suggest that increased NGF signaling disrupts homeostasis in the decidual region, leading to elevated levels of lipid peroxidation and inflammation, which in turn trigger preterm birth.

Materials and methods

NGF Recombinant protein purification

NGF (Uniprot P01139, residues 122–241) cDNA was inserted into the pET-28a vector featuring a hexahistidine (6xHis) tag and an MBP tag at its N-terminus. Roseta E.coli transformed with each plasmid were cultivated in LB media and induced with 3 mM isopropyl-β-D-thiogalactoside (IPTG) upon reaching a density of 0.6 OD. The bacterial culture was incubated at 16 °C for 12–16 h. Pelleted cells were resuspended and sonicated in lysis buffer containing 250 mM NaCl and 30 mM HEPES (PH 7.5). His-tagged MBP-NGF was purified using Ni2+-NTA resin and eluted with an imidazole-containing buffer. The eluted proteins underwent further purification via a Superose 6 increase column (GE Healthcare, 29091596) and were collected after passage through an endotoxin removal affinity column (Thermo Scientific, 88274). Protein concentration was determined using a BCA assay kit (Thermo Scientific, 23227).

Mice and treatments

The evaluation of pregnancy events was conducted as previously described [24]. C57BL/6J female mice aged 8–12 weeks were mated with reproductively competent wild-type male mice to induce pregnancy (the presence of a vaginal plug was considered day 1 of pregnancy). On the morning of day 16 of pregnancy at 10:00, intraperitoneal injections of MBP-NGF (containing 25 µg of NGF, a dose slightly elevated compared to literature due to increased maternal body weight, expanded blood volume, and accelerated drug clearance during late gestation) were administered to mice [19], while the control group received equivalent recombinant MBP-tag protein. For erastin treatment (Selleck, S7242), mice received intraperitoneal injections of 20 mg/kg of erastin once daily from day 16 to day 18 in the morning, while the control group received an equivalent volume of the vehicle without erastin. Gestational durations of less than 19.5 days were classified as preterm, whereas deliveries occurring between 19.5 and 20 days were designated as full-term. On gestational day 17, the mice were anesthetized with isoflurane inhalation. Both tissue and blood samples were collected and subsequently utilized for molecular biology experiments.

In situ hybridization

Probes for Tpbpα, Prl3c1, Prl3b1, and Prl7b1 were generated using the SP6/T7 Transcription Kit (Roche, 11175025910). Linear template DNA was transcribed into digoxigenin-labeled RNA using the digoxigenin-RNA Labeling Mix (Roche, 11277073910). After transcription, 1 µL of DNaseI (Biolabs, M0303S) was added to digest the DNA template. Subsequently, RNA was precipitated overnight at -80 °C by adding 2 µL of 0.5 M EDTA, 2.5 µL of 4 M LiCl, and 75 µL of chilled ethanol. The purified RNA probe was then obtained. Frozen sections were processed onto the same slide and hybridized with probes labeled with digoxigenin. The signals were observed under bright-field microscopy using a panoramic scanner (OLYMPUS, VS120). The probe sequence is displayed in the supplementary material.

Quantitative qRT-PCR

Frozen decidua and placenta samples were prepared for total RNA extraction with TRIzol reagent (Invitrogen, 15596-018). The specific experimental procedures were conducted in accordance with previous research [25]. The RNA was reverse transcribed using a reverse transcription kit (Vazyme, R211-01) following the protocol, and the cDNA was diluted five-fold for quantitative PCR reaction using the cDNA dilution directly. Quantitative real-time PCR was performed using SYBR Green Master Mix (Vazyme, Q311-02) on a Real-Time PCR system (Light Cycler 96), and the specific primers are listed in the supplementary material. The relative mRNA levels of each gene were calculated using the reference genes Rpl7 and β-Actin as internal controls.

Immunofluorescence

Frozen Section (12 μm) from both the control and experimental groups were placed on the same slide and fixed with 4% paraformaldehyde (PFA) (Sigma, 158127-500) for 10 min. The sections were then washed with 1× PBST (PBS + 0.1% Triton X-100) and blocked with a 5% bovine serum albumin (BSA) solution. Subsequently, primary antibodies were applied at 4 °C overnight, followed by visualization using the corresponding secondary antibodies conjugated with Alexa 488 or Alexa 594. Nuclei were stained using DAPI (Solarbio, C0065, 5 µg/mL). Images were captured using a panoramic fluorescence microscope (OLYMPUS, VS120) and processed using OlyVIA software. The antibodies used in this study are listed in the supplementary material.

Immunohistochemistry

For immunohistochemistry (IHC). After fixation of tissue sections with 4% PFA, antigen retrieval was performed using EDTA. Endogenous peroxidase activity was blocked using an endogenous peroxidase blocking solution (Beyotime, P0100A). Tissues were then treated with 5% BSA for 60 min to prevent nonspecific binding, followed by incubation with the primary antibodies and secondary antibodies sequentially. The antibodies used in this study are listed in the supplementary material. Immunostaining visualization was performed using the DAB substrate kit (Zhongshan Golden Bridge Biotechnology, ZLI-9018). Images were captured and analyzed using bright-field panoramic microscopy (OLYMPUS, VS120).

Measurement of serum PGF2α, P4 and E2 levels

On gestational day 17, serum samples were collected, and the concentrations of prostaglandin F2α (PGF2α), progesterone (P4) and estradiol (E2) were quantitatively measured using enzyme-linked immunosorbent assay (ELISA) kits (Cayman, 16010, 582601 and 501890, respectively).

FACS-based lipid peroxidation assay

Dissected the decidual tissue, rinsed it with pre-cooled PBS, minced it into a slurry, and moved it to accumax (Sigma, A7089) for cell digestion for 30 min at 37 °C. After terminating the digestion, the cells were passed through a 200-mesh sieve, washed with PBS, and resuspended. To detect lipid peroxidation, cells were resuspended with PBS containing 5 µM C-11 BODIPY 581/591 dye (Invitrogen, D3861) and incubated at room temperature for 30 min, and data were processed using FlowJo software.

Immunoblotting

Proteins from tissues were extracted using RIPA buffer (Beyotime, P0013B) supplemented with 1% protease inhibitor cocktail (Roche, 04693132001) or phosphatase inhibitor cocktail (Roche, 04906845001). The protein concentration was quantified using the Pierce BCA Protein Assay Kit (Thermo, 23227). Total proteins were separated via 10% or 15% SDS-PAGE gel electrophoresis and subsequently transferred onto PVDF membranes (Millipore, ISEQ00011). Following the transfer, membranes were blocked with 5% skim milk at room temperature for 1 h, then incubated with specific primary antibodies and appropriate secondary antibodies. β-Actin served as a loading control. The antibodies used in this study are listed in the supplementary material. Protein bands were visualized using the ECL substrate kit (Thermo, 32106) and detected with chemiluminescence imaging system (Tanon 5200). Quantify the density of protein bands using ImageJ software.

Measurement of MDA and 12(S)-HETE levels

Measurement method of MDA was modified according to previous research [26]. Extract the protein from the decidual tissue and quantify the protein concentration using the BCA assay. Then, adjust the protein concentration to equivalent levels. Subsequently, equal volumes were taken for MDA determination according to the serum measurement method. 100 µL of serum was added to 325 µL of 1-methyl-2-phenylindole (Shanghai Yuanye Bio-Technology, S46375), which was subsequently dissolved in a mixture of acetonitrile (Sigma-Aldrich, 34851) and methanol (3:1, V/V), achieving a final concentration of 10 mM 1-methyl-2-phenylindole. The reaction was initiated by the addition of 75 µL of 37% hydrochloric acid. After incubating the reaction mixture at 70 °C for 45 min, the samples were centrifuged (10 min, 15,000 g, 4 °C) and the supernatant was collected. The absorbance at 595 nm was measured, and the MDA concentration was determined using 1,1,3,3-tetramethoxypropane (Sigma-Aldrich, 108383) as the MDA standard. The measurement of 12(S)-HETE was conducted in accordance with the protocol provided in the reagent kit (Abcam, ab133034).

Transmission electron microscopy

Fresh mouse decidual tissue was fixed in 2.5% glutaraldehyde/0.1 M phosphate buffer (PH 7.4) for 5 min. Subsequently, the tissue was sectioned into 1 mm² slices. These tissue blocks were then fixed at room temperature for 2 h and treated overnight at 4 °C. The regions of interest were located on the cross-section of the tissue. Samples were rinsed with 0.1 M PBS and fixed with 1% osmium acid fixative (PH 7.4) for 2 h. After a 30 min rinse with phosphate buffer, the samples underwent a series of ethanol treatments, followed by gradient dehydration using 100% acetone. The samples were then soaked and embedded in epoxy resin, and polymerized using an automatic tissue processor (Leica, EM TP). The tissue was subsequently cut into semi-thin sections of 5 μm thickness using an ultramicrotome (Leica, EM UC7) and dried for 20 min. These sections were stained with toluidine blue (Sigma, 89640) and examined under a light microscope. Ultrathin sections were then cut to a thickness of 50 nm using an ultramicrotome, stained with saturated aqueous uranyl acetate and lead citrate solution for 30 min in a staining machine (Leica, EM AC20), and subsequently photographed and analyzed using a Thermo Fisher transmission electron microscope (Talos F200C).

Statistical analysis

Statistical analyses were performed using GraphPad Prism 9.0 (GraphPad Software Inc., La Jolla, CA, USA). Data are presented as the mean ± SD. Firstly, verify if the data follows a normal distribution. If it does, significant differences were calculated by two-tailed Student’s t-test, if not, significant differences were calculated by Mann-Whitney test. P value < 0.05 was considered statistically significant.

Results

Intraperitoneal injection of recombinant NGF protein at gestational day 16 induces preterm birth in mice

To investigate the potential role of NGF in late pregnancy, we employed a single-cell database (PRJNA907416) to evaluate the expression of NGF and its receptors during late-stage pregnancy in mice [27]. Our analysis revealed that as pregnancy progressed, the expression levels of NGF, along with its receptors p75NTR (Ngfr in rodents) and TrkA (Ntrk1 in rodents), were markedly elevated in decidual cells on gestational day 19 compared to day 16. (Fig. S1A-C). Furthermore, we extended our analysis to a dataset (GSE200289) derived from an E. coli-induced preterm birth mouse model [28], we observed a significant elevation in the expression of Ngf and its associated receptors in decidual cells compared to the control group (Fig. S1D-F). These results suggest that NGF signaling may be involved in the initiation of labor.

In the present study, following the administration of a 25 µg dose of NGF via intraperitoneal injection on the morning of day 16, approximately 64% of the mice exhibited preterm labor, defined as giving birth prior to day 19. Notably, the majority of mice that underwent premature delivery were concentrated in the evening of day 17. This outcome had adverse consequences, as the neonatal pups faced significant challenges in their survival. In contrast, the control group, which received an equivalent dose of MBP, did not exhibit any instances of preterm labor. (Fig. 1A-C).

Fig. 1

Intraperitoneal injection of recombinant NGF protein at gestational day 16 induces preterm birth in mice. (A) Pregnant C57BL/6J mice were intraperitoneal injected with recombinant NGF protein (25 µg NGF, MBP group mice were given an equal MBP protein content) on day 16. (B) Gestational lengths (n = 9 for MBP group, n = 11 for NGF group). (C) Preterm birth rates. (D) qPCR analysis of genes encoding Ngfr related binding proteins in maternal decidua on day 17, (n = 6 mice per group). (E) Western blotting of Ngfr and Pan-Trk in maternal decidua on day 17, (n = 4 mice per group). (F) Quantitative statistics of Western blotting. (G) IHC of NGFR in maternal-fetal interface on day 17. LE, luminal epithelium. Scale bars, 200 μm. Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001. In (B), by Mann-Whitney test. In (D) and (F), by two-tailed Student’s t-test

Moreover, we investigated the effects of NGF treatment on its associated receptors. Our results demonstrated a significant upregulation in the expression levels of both NGFR and TRK receptors following NGF treatment in the maternal decidual zone (Fig. 1E-G). The biological functions of NGFR are mediated through its interactions with various binding proteins. For instance, NGFR forms a complex with brain expressed X-linked protein 3 (BEX3), which is crucial for NGF-dependent regulation of neuronal apoptosis and survival [29, 30]. Intriguingly, our findings revealed a significant decrease in both Bex3 mRNA and protein levels within the mouse decidua after NGF treatment. It is worth noting that a recent study has shown that BEX3, as a pseudo-substrate for Cul2-FEM1B ubiquitination, plays a crucial role in the redox homeostasis, and knockout of BEX protein leads to a significant increase in ROS levels [31]. Unexpectedly, other NGFR -binding factors, such as Maged1 and Sort1 [32, 33], did not show significant alterations (Fig. 1D and E). Taken together, our findings suggest that elevated NGF could activate its associated receptors NGFR and NTRK family members, thereby potentially inducing preterm birth in mice.

Mice treated with NGF displayed normal structure and function of placenta and ovary

Abnormal fetoplacental growth and hormone changes are factors that contribute to preterm birth, we then detected the potential effects of NGF treatment on placenta and ovary [5, 34]. CDX2 plays a crucial role in regulating placental cell proliferation, turnover, and differentiation into outer layer cells, ensuring proper placental development and function. Aberrant CDX2 expression or function can disrupt placental function, with implications for fetal and maternal well-being [35, 36]. Our evaluation revealed that no significant alterations in the expression and distribution of CDX2 protein in both the maternal decidua and the sponge trophoblast zone of placenta between the NGF treatment and control groups (Fig. 2A). Prl3b1 and Prl7b1, members of the PRL (Prolactin) peptide hormone family, are primarily expressed in placental trophoblast cells and are thought to regulate placental development and function [37]. Furthermore, Gcm1 and Tpbpα are closely associated with the differentiation and function of placental cells, contributing to the regulation of placental growth and development [38, 39]. However, our in situ hybridization and qPCR analyses targeting specific placental tissues demonstrated no impact of NGF treatment on placental development (Fig. 2B and C). This was evidenced by the absence of significant changes in placental structure, morphology, and key transcription factor levels when compared to the control group. Monocarboxylate transporter 1 (MCT1) and monocarboxylate transporter 4 (MCT4) are critical for maintaining substance exchange and metabolic balance between the fetus and placenta. Impaired MCT1 and MCT4 function can disrupt lactate metabolism, adversely affecting fetal growth and development [40, 41]. Nonetheless, analyses of the labyrinth zone indicated that NGF treatment did not affect the expression of MCT1 and MCT4, suggesting normal substance exchange between mother and fetus (Fig. 2D-F).

Fig. 2

Mice treated with NGF displayed normal placental structure and function. (A) IF of CDX2 and CK8 in maternal-fetal interface on day 17. Scale bars, 500 μm. Dec, decidua. Sp, spongiotrophoblast. Lb, labyrinth. (B) qPCR analysis of spongiotrophoblast related genes in placenta on day 17, (n = 6 mice per group). (C) In situ hybridization of Prl3b1 (sponge trophoblast cells) and Prl7b1 (trophoblast giant cells) in spongiotrophoblast on day 17. Scale bars, 500 μm. (D) IF of MCT1 and MCT4 in labyrinth on day 17. Scale bars, 500 μm. (E) Western blotting of MCT1 and MCT4 in maternal decidua on day 17, (n = 4 mice per group). (F) Quantitative statistics of Western blotting. In (B) and (F), data are presented as mean ± SD, ns represent no significant statistical difference, by two-tailed Student’s t-test

The withdrawal of P4 is a critical factor initiating parturition [42]. Therefore, we investigated the impact of NGF on ovarian function. Immunohistochemical results demonstrated that the protein levels of key enzymes involved in P4 synthesis, CYP11A1 and HSD3B1, in the ovary were not affected by NGF (Fig. S2A). Additionally, we quantified the levels of serum estrogen (E2) and progesterone (P4) on Day 17. ELISA results revealed that the levels of P4 and E2 in the serum following NGF treatment were comparable with the control group (Fig. S2B and C). This finding indicates that NGF does not induce changes in ovarian function.

Mice treated with NGF did not show signs of senescence and exhibited a normal mTOR signaling pathway in the maternal decidua

Prl8a2 and Prl3c1 are indicators of decidual terminal differentiation, representing the degree of decidual differentiation [43]. Interestingly, NGF treatment caused significant upregulation of decidual differentiation markers in late pregnancy (Fig. 3A). In a comprehensive study exploring the genetic and gene-environment influences during pregnancy, targeted loss of Trp53 in the uterine tissue has been found to result in spontaneous preterm birth in around 50% of mice [6]. The absence of Trp53 induces pronounced decidual senescence within the maternal decidua, leading to dysregulation of critical signaling pathways like mTOR and AMPK [34, 44, 45]. In investigating the effects of NGF administration on the decidual tissue of preterm birth mice, our investigations using β-SA-gal and γ-H2AX assays yielded no evidence of decidual senescence (Fig. 3B). Furthermore, the mRNA expression levels of genes associated with senescence-related pathways, such as Trp53, Cdkn1a, and Cdkn2a, remained unchanged within the maternal decidua in NGF group compared to the MBP group (Fig. 3C). Additionally, analysis via Western blotting of molecules implicated in classical senescence pathways like AMPK and mTOR showed no alterations in protein levels induced by NGF treatment (Fig. 3D and E). As expected, immunofluorescence staining of Ki67 on day 17 showed no noticeable anomalies in cellular proliferation within the decidual region. It is notable that minimal cell proliferation was observed solely around blood vessels. (Fig. 3F). Our findings indicate that NGF induced preterm labor is not mediated by decidual senescence.

Fig. 3

Mice treated with NGF exhibited enhanced decidualization without impacting senescence and exhibited a normal mTOR signaling pathway in the maternal decidua. (A) In situ hybridization of Prl8a2 and Prl3c1 (terminal differentiation marker of decidua) on day 17. Scale bars, 500 μm. (B) β–SA-gal chemical stain and IHC of γ-H2AX in maternal-fetal interface on day 17. Scale bars, 500 μm. (C) qPCR analysis of cell senescence related genes in maternal decidua on day 17 (n = 6 mice per group). (D) Western blotting of important molecules of cell senescence pathway in maternal decidua on day 17 (n = 4 mice per group). (E) Quantitative statistics of Western blotting. (F) IF of Ki67 in maternal decidua on day 17. Scale bars, 100 μm. BV, blood vessel. In (C) and (E), data are presented as mean ± SD, ns represent no significant statistical difference, by two-tailed Student’s t-test

NGF treatment activates the p-STAT3 pathway and promotes the expression of proteins associated with labor initiation in the maternal decidua

Cyclooxygenase (COX) is the enzyme that catalyzes the synthesis of prostaglandins (PGs) from aromatic acid compounds. COX1 and COX2 are crucial enzymes in prostaglandin synthesis, converting arachidonic acid (AA) into the epoxy intermediate prostaglandin G2 (PGH2), which is further transformed into various prostaglandins, ultimately producing PGF2α [46]. As expected, COX1 protein levels significantly increased following NGF treatment, with COX1 highly enriched in epithelial cells. COX2, which is significantly expressed under inflammatory and other stimulatory conditions, is regarded as a key marker for the initiation of labor. COX2 expression was markedly elevated post-NGF treatment and localized at the both maternal decidual and sponge trophoblast zone (Fig. 4A-D). Initiation of labor is regulated by decidual PGF2α, with significantly elevated levels observed in NGF-treated mice on day 17 compared with the control group (Fig. 4E). pSTAT3 is a pivotal mediator of inflammatory responses, and Western blotting results of the decidual tissue showed that NGF-induced pSTAT3 expression within the maternal decidua in mice was significantly higher compared to the control group, especially the phosphorylation of Tyr705 site. However, the ratio of p-STAT3 to total STAT3 showed no significant difference compared to the control group following NGF treatment, which was attributed to the NGF-induced moderate upregulation of total STAT3 levels. (Fig. 4F and G). Moreover, IHC results showed that the phosphorylation level at Tyr705 site increased significantly in the NGF-treated group compared to the control group in the decidual area (Fig. 4H). We detected mRNA expression of Cxcls and the ILs family. As expected, these pro-inflammatory factors were significantly upregulated after treatment with NGF, indicating the occurrence of pSTAT3 mediated inflammatory response (Fig. 4I).

Fig. 4

NGF treatment activates the p-STAT3 pathway and promotes the expression of proteins associated with labor initiation in the maternal decidua. (A) IHC of COX1 in maternal decidua on day 17. Scale bars, 200 μm. LE, luminal epithelium. (B) IF of COX2 and CK8 in maternal-fetal interface on day 17. Scale bars, 500 μm. (C) Western blotting of COX1 and COX2 in decidua on day 17 (n = 4 mice per group). (D) Quantitative statistics of Western blotting. (E) PGF2α levels in serum on day 17 (n = 10 for MBP group, n = 11 for NGF group). (F) Western blotting of p-STAT3 and STAT3 in decidua on day 17 (n = 4 mice per group). (G) Quantitative statistics of Western blotting. (H) IHC of p-STAT3 Tyr705 in maternal decidua on day 17. Scale bars, 500 μm. Dec, decidua. Sp, spongiotrophoblast. Lb, labyrinth. (I) qPCR analysis of inflammation related genes in maternal decidua on day 17 (n = 5 mice per group). In (D), (E), (G) and (I), data are presented as mean ± SD. *p < 0.05, **p < 0.01, ****p < 0.0001, ns: not significant, by two-tailed unpaired Student’s t-test

NGF treatment activates inflammatory response and lipid peroxidation pathway in the maternal decidua

To further investigate the mechanism of NGF-induced preterm birth, we isolated the decidua region from day 17 and performed RNA-seq, which enriched significantly upregulated and downregulated genes (Fig. 5A and B). In addition, gene set enrichment analysis (GSEA) and Gene Ontology (GO) analysis showed that inflammatory-associated genes were significantly enriched by NGF treatment (Fig. 5C-D). GO results revealed significant enrichment of acute inflammatory response pathways, and associated inflammatory factors such as Nlrp1a, Cxcls and Ccls were markedly activated (Fig. 5E). We also validated the relevant pro-inflammatory cytokines including Ccls, Cxcls, Nlrp3 and Tnfrsf1a through qPCR, showing a significant increase in mRNA levels in NGF group compared to the control group (Fig. 5F).

Fig. 5

NGF treatment activates inflammatory response and lipid peroxidation pathway in the maternal decidua. (A) Heatmap of normalized gene expression levels of mice decidua treated with MBP or NGF. Blue indicates down-regulation of genes expression; Red indicates up-regulation of genes expression. (B) Volcano plot of differentially expressed genes in decidua. Blue and red dots represent genes that are differentially expressed (p < 0.05) between MBP and NGF group. Genes not significantly differentially expressed are shown in black. (C) GSEA plot showing the enrichment of acute inflammatory response-related genes. NES, normalized enrichment score. (D) Gene Ontology functional analysis for differentially expressed genes. (E) Heatmap of inflammation associated genes differentially expressed between MBP and NGF group. (F) qPCR analysis of inflammation related genes in maternal decidua on day 17 (n = 6 mice per group). (G), (H) and (I) GSEA plot showing the enrichment of ROS metabolic process, Unsaturated fatty acid metabolic process and Glutathione metabolism-related genes respectively in NGF group compared with MBP group. NES, normalized enrichment score. (J) Heatmap of lipid peroxidation associated genes differentially expressed between MBP and NGF group. (K) qPCR analysis of lipid peroxidation related genes in maternal decidua on day 17 (n = 6 mice per group). In (F) and (K), data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001 by two-tailed unpaired Student’s t-test

In addition to the enrichment of acute inflammation pathways, we also noted that the results of the GO enrichment analysis showed significant alteration in the pathways of reactive oxygen species, the transport of long-chain fatty acids, and the synthetic metabolism of glutamine (Fig. 5D). The results of GSEA showed abnormal activation of genes related to the ROS pathway, unsaturated fatty acid metabolism, and glutamine metabolism (Fig. 5G-I). This finding implied that the maternal decidua might have encountered disruptions in the homeostasis of lipid oxidation following NGF treatment. Redox homeostasis plays a crucial role in normal cellular physiology. Excess lipid peroxides disrupt this balance, contributing to various diseases and acting as key drivers of ferroptosis [47]. Reactive oxygen species nonspecifically attack lipids containing carbon-carbon double bonds, producing lipid hydroperoxides (LOOH), a process known as lipid peroxidation [25]. As anticipated, our analysis has elucidated the collaborative dynamics of gene expression alterations associated with lipid peroxidation. Notably, the expression of Alox12 (Arachidonate 12-lipoxygenase) was upregulated, reflecting its role in catalyzing the oxidation of fatty acids. Concurrently, Trf (Transferrin), a key regulator in iron metabolism, was downregulated. Gpx4 (Glutathione Peroxidase 4) expression was reduced, which underscores its important function in catalyzing the reaction between reduced glutathione (GSH) and peroxides, thereby protecting cells from oxidative stress-induced damage (Fig. 5J). To corroborate these findings, we conducted mRNA level analyses focusing on genes such as Gpx4, Keap1, and Trf, which are critical in the regulation of lipid peroxidation. The expression levels of a multitude of molecules were found to be dysregulated (Fig. 5K). These collective data suggest that following NGF treatment, the maternal decidua exhibits pathological traits, including inflammation and lipid peroxidation.

NGF treatment leads to an elevation in levels of lipid peroxidation in the maternal decidua

We assessed the protein levels of GPX4, observing significant reductions in protein levels using the tissue of maternal decidua, Moreover, IHC results showed that the positive signal of GPX4 in the decidua is similarly significantly reduced in the NGF group compared to the control group (Fig. 6A-C). Malondialdehyde (MDA), a metabolite of lipid peroxidation, showed a marked increase in both whole blood and decidual tissue following NGF treatment compared to controls (Fig. 6D and E). 12-HETE (12-Hydroxyeicosatetraenoic acid) indirectly promotes lipid peroxidation by enhancing the activity of lipid peroxidases and promoting the generation of free radicals [48]. Additionally, serum levels of 12-HETE were significantly elevated in the NGF group compared with the control group (Fig. 6F). Lipid peroxidation induces changes in mitochondrial function and morphology, such as loss of membrane integrity and cristae reduction. The electron microscopy results indicated that NGF treatment led to a significant decrease in mitochondrial cristae and compromised membrane integrity in the decidual zone of day 17 mice compared to controls (Fig. 6G). BODIPY 581/591 C11, a fluorescence ratio probe for lipid oxidation, allows for the visualization of lipid peroxidation through flow cytometry [49]. Our results demonstrated increased levels of lipid reactive oxygen species in primary decidual cells dissociated on day 17, indicating that NGF treatment elevates lipid peroxidation at the maternal-fetal interface (Fig. 6H and I). These results have confirmed that NGF treatment can lead to the occurrence of lipid peroxidation in the maternal decidua.

Fig. 6

NGF treatment leads to an elevation in levels of lipid peroxidation in the maternal decidua. (A) Western blotting of GPX4 in maternal decidua on day 17 (n = 4 mice per group). (B) Quantitative statistics of Western blotting. (C) IHC of GPX4 in maternal decidua on day 17. Scale bars, 500 μm. Dec, decidua. (D), (E) MDA levels in decidua (n = 10 for MBP group, n = 10 for NGF group) and serum (n = 11 for MBP group, n = 11 for NGF group) respectively on day 17. (F) 12-HETE levels in serum on day 17 (n = 14 for MBP group, n = 13 for NGF group). (G) Representative transmission EM images of the organelle in the decidual on day 17. Mito, mitochondrion. Scale bar, 500 nm. (H) Lipid peroxidation levels in decidual cells isolated from MBP and NGF treated mice on day 17 (n = 4 mice per group). (I) Quantitative statistics of lipid peroxidation levels. In (B), (D), (E), (F) and (I), Data are presented as mean ± SD. *p < 0.05, **p < 0.01, by two-tailed unpaired Student’s t-test

Intraperitoneal injection of erastin induces preterm birth in mice

Given the remarkable elevation of lipid peroxidation indicators in the maternal decidua following NGF treatment, and considering the mounting evidence linking lipid peroxidation as an indicator closely associated with preterm birth, we decided to employ lipid peroxidation inducers in mice on late stagy of pregnancy. On day 16 of murine pregnancy, we administered the erastin (20 mg/kg) intraperitoneally for three consecutive days from day 16 of pregnancy (Fig. 7A). As expected, about half of the mice began delivering by the afternoon of day 18, with a few delivering by the morning of day 19 (Fig. 7B and C), the mortality rate of the offspring after erastin treatment was 73.3% (Fig. 7D).

Fig. 7

Intraperitoneal injection of erastin induces preterm birth in mice. (A) Pregnant C57BL/6J mice were intraperitoneal injected with erastin (20 mg/kg) intraperitoneally for three consecutive days to induce lipid peroxidation. (B) Gestational lengths (n = 9 for Vehicle group, n = 6 for Erastin group). (C) Preterm birth rates. (D) Neonatal mortality. (E) qPCR analysis of lipid peroxidation related genes in maternal decidua on day 17, (n = 6 mice per group). (F) IHC of GPX4 in maternal decidua on day 17. Scale bars, 500 μm. Dec, decidua. (G) Western blotting of GPX4 in maternal decidua on day 17 (n = 4 mice per group). (H) Quantitative statistics of Western blotting. (I) MDA levels in the decidua on day 17 (n = 12 for Vehicle group, n = 11 for Erastin group). (J) Lipid peroxidation levels in decidual cells isolated from erastin and vehicle treated mice on day 17 (n = 4 mice per group). (K) Quantitative statistics of lipid peroxidation levels. (L) qPCR analysis of genes encoding Ngf and related receptors in maternal decidua on day 17, (n = 6 mice per group). Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ****p < 0.0001, ns: not significant. In (B), by Mann-Whitney test. In (E), (H), (I), (K) and (L), by two-tailed unpaired Student’s t-test

We conducted a comprehensive study focusing on placental structure and decidual senescence after erastin administration. Consistent with the findings from NGF-induced preterm labor in mice, erastin treatment similarly showed no alterations in the structural and functional markers of the placenta, including Cdx2, Prl3b1, Tpbpα, and Prl7b1, at day 17. Specifically, the structure and function of the spongiotrophoblast layer remained unchanged (Fig. S3A, E and F). Similarly, the markers of the labyrinth zone, MCT1 and MCT4, exhibited no significant changes following erastin treatment (Fig. S3B-D). Additionally, we assessed decidual senescence, with β-SA-gal and γ-H2AX assays revealing no significant signs of senescence in the decidua after erastin exposure. (Fig. S4A). Consistent with these findings, no significant changes were observed at the molecular level in senescence-related markers, including such as Trp53, Cdkn1a, and Cdkn2a, within the decidua (Fig. S4B).

To further investigate the alterations in lipid peroxidation levels within the decidual region, we assessed the expression of relative genes and found that erastin treatment led to a reduction in the levels of genes involved in the lipid peroxidation cascade within the decidua, including Gpx4 and Trf. Additionally, a significant decline in GPX4 protein expression was observed in the decidual tissue and decidual zone (Fig. 7E-H). Further analysis involved the quantification of MDA levels, which revealed that the MDA content in the decidua was substantially higher in the erastin-treated group compared to the vehicle control group (Fig. 7I). The results obtained from the BODIPY 581/591 C11 assay also indicated that the level of lipid peroxidation in primary cells isolated from decidua was significantly elevated in the erastin-treated group relative to the control group (Fig. 7J and K). These studies suggest that the augmented decidual lipid peroxidation levels play a pivotal role in the induction of preterm labor. We examined key markers related to the initiation of parturition, revealing a significant upregulation of COX1 and COX2 (Fig. S5 A-D). Additionally, pSTAT3 levels in decidual tissue were markedly elevated at both the Tyr705 and Ser727 phosphorylation sites following erastin treatment (Fig. S5 E-H). We sought to determine whether excessive lipid peroxidation induces alterations in the expression of NGF and its receptors within the maternal decidua. Endogenous NGF expression is increased in inflammatory conditions, and co-administration of NGF-neutralizing antibodies during LPS treatment markedly decreases the release of the pro-inflammatory cytokine IL-10 [11]. Notably, erastin treatment significantly upregulated the expression of the Ngfr in the maternal decidua, accompanied by a modest upward trend in Ngf levels, though the change in Ngf did not reach statistical significance (Fig. 7L). The elevation of NGF is more likely similar to the upregulation seen in the IL family or Cxcl family, both of which are significantly increased following an inflammatory response, suggesting a pro-inflammatory role.

In summary, our findings suggest that increased NGF signaling disrupts homeostasis in the maternal decidual, leading to elevated levels of lipid peroxidation and inflammation, which in turn trigger preterm birth.

Discussion

NGF levels during pregnancy show a 1.7-fold increase compared to non-pregnant women, indicating its important role in the process of pregnancy [50]. This work has investigated the pivotal role of elevated NGF in modulating physiological processes during late pregnancy, particularly its involvement in preterm birth in mice. The significant premature birth observed in the NGF-treated group emphasizes the detrimental effects of aberrant NGF signaling, suggesting that while NGF is critical for the initiation of labor, its overexpression during late pregnancy can disrupt the delicate balance required for successful gestation. Through analyzing spatial transcriptome data on mouse pregnancy, we obtained spatiotemporal expression patterns of NGF and its receptors during mid-pregnancy. By localizing the decidual marker Bmp2 and the spongiotrophoblast marker Slc16a1, we found that the receptors NTRK1 and NTRK2 for NGF are mainly expressed in the maternal decidua, as was previously reported [51]. Furthermore, the single-cell analysis supports a marked upregulation of NGF and its receptors, NGFR and NTRK1, in decidual cells as pregnancy progressed [27]. Those findings imply that NGF signaling is not merely a bystander but plays an active role in the processes leading to labor initiation. The association between NGF and the inflammatory responses observed further corroborates the hypothesis that NGF is integral to the physiological changes that precede parturition. Specifically, our results showed that NGF treatment led to increased expression of COX2 and PGF2α, both of which are essential mediators in the labor process [6, 52]. The elevation of these inflammatory markers suggests that NGF may instigate a cascade of events that culminates in labor, potentially through the activation of inflammatory pathways.

A previous study reported that following overexpression of NGF in mouse ovaries, a decrease in P4 levels during the estrous cycle and impaired response to hCG-like activity were observed [53]. The pleiotropic effects of the NGF pathway, particularly its impact on ovarian function, prompt us to first investigate whether there are alterations in ovarian function following administration of NGF. It is worth noting that in our study, NGF treatment did not decrease serum P4 levels nor affect the key P4 synthesis enzymes in the ovary, which indicates that the increase in NGF levels leading to preterm birth is not caused by disorder in ovarian function. It may be due to differences in the doses and time frame of administration used in our study, as the treatment of NGF affecting ovarian function was administered to mice after ovulation [54]. Meanwhile, our finding suggests that the mechanisms by which NGF induces labor are distinct from those leading to placental dysfunction [55, 56]. The lack of significant changes in markers like CDX2 and MCT1/4 reinforces this notion, indicating that NGF’s influence may primarily target the decidua and its associated signaling pathways rather than directly affecting placental morphology or function. Therefore, it is evident that the induction of preterm birth by elevated NGF levels in late pregnancy is not a result of alterations in ovarian or placental function, but rather stems from physiological changes within the decidual region.

Decidual senescence is one of the key factors contributing to preterm birth [6, 45], an unexpected finding in our study was the absence of decidual senescence in NGF-treated mice, despite the significant upregulation of decidual differentiation markers. This suggests that NGF’s role in promoting preterm birth may not be mediated through senescence pathways. The absence of alterations in mTOR signaling, a critical pathway often associated with cellular aging [44], further substantiates this interpretation. Instead, our results suggest that NGF may activate specific inflammatory pathways, particularly through the p-STAT3 signaling route, which is known to mediate responses to inflammatory stimuli [7]. In examining the underlying mechanisms of NGF-induced preterm birth, our RNA-seq analysis indicated a robust activation of inflammatory and lipid peroxidation pathways in the maternal decidua. This is particularly noteworthy, as lipid peroxidation has been recognized as a critical factor in various pathological states and has also been suspected to play a role in triggering preterm birth [57, 58].

Our recent study showed that specific deletion of uterine GPX4 increases uterine lipid peroxidation levels, resulting in implantation failure and infertility [25]. The aforementioned results indicate that the role of lipid peroxidation during late pregnancy warrants further investigation. In our transcriptome data, we significantly enriched unsaturated fatty acid synthesis, glutathione synthesis, and ROS pathways. Polyunsaturated fatty acids on cell membranes, can be oxidized via non-enzymatic pathways (ROS induced by various oxidative stresses) or enzymatic pathways (such as Lipoxygenase, LOXs), forming lipid hydroperoxides [47, 59]. Simultaneously, genes regulating the level of lipid peroxidation exhibited coordinated changes. This provides further evidence that NGF treatment indeed results in an elevation of lipid peroxidation levels in the maternal decidua region. Regarding the relationship between lipid peroxidation and preterm birth, studies have shown significantly higher MDA levels in PTB umbilical cord blood compared to TB [60]. In the PTB group, low levels of antioxidants SOD and CAT and total protein were observed, while malondialdehyde (a byproduct of lipid peroxidation) was higher. Compared to TB, PTB urine [61] and amniotic fluid [62] had higher levels of 8-isoprostanes (an F2-isoprostane). In this study, we observed that NGF treatment significantly increased levels of MDA in serum and protein samples of decidual tissue, and BODIPY 581/591 C11 levels, the most direct indicator of lipid peroxidation in the decidual region, were significantly elevated, accompanied by downregulated expression of the key antioxidant enzyme GPX4. This suggests that the NGF-induced lipid peroxidation may facilitate an environment conducive to inflammatory response, further exacerbating the risk of premature delivery.

Interestingly, we observed that administration of erastin, a known inducer of lipid peroxidation, also resulted in preterm birth, reinforcing our hypothesis that lipid peroxidation plays a critical role in labor initiation. Our findings revealed that erastin administration precipitated an upsurge in lipid peroxidation within the maternal decidua, eventually culminating in preterm birth.

This observation prompts further investigation into the mechanisms by which NGF mediates the augmentation of lipid peroxidation levels. NGF treatment alone in immune cells causes significant release of superoxides including cytochrome C [63, 64]. The release of superoxides induced by NGF treatment can oxidize abnormal lipid metabolism products, disrupting lipid redox homeostasis and inducing pathological states like lipid peroxidation. This may provide a reasonable explanation for why NGF can induce an increase in the level of lipid peroxidation in the maternal decidua, however, it still needs to be clarified how NGF-induced ROS activation leads to changes in the expression of key lipid peroxidation factors such as GPX4 and ALOX12, whether through the activation of oxidative stress-sensitive transcription factors or inflammation-related signaling pathways. In our previous research, treatment with lipid peroxidation scavengers Liproxstatin-1 in mice was unable to rescue the implantation failure, as STAT3 underwent carbonylation in the high lipid peroxidation environment, leading to a loss of its function [25]. STAT3 is not only an important factor in regulating embryo implantation, but in our earlier studies, we also found that STAT3 is crucial for the initiation of labor. Therefore, the treatment with antioxidant drugs may not eliminate the premature birth phenotype caused by lipid peroxidation induced by NGF.

In this study, NGF was shown to induce strong inflammatory responses and lipid peroxidation. Regarding the relationship between inflammation and lipid peroxidation, treatment with octanal has been shown to induce lipid peroxidation and increase LPS-induced IL-1β production. However, the absence of NLRP3 reversed the contribution of octanal to oxidative stress in macrophages [65]. Conversely, supplementation with CoQ10 enhances mitochondrial function, reduces excessive lipid peroxidation, and ultimately inhibits the formation of NLRP3 inflammasomes and cell death in ETFDH-mutated lymphoblastoid cells [66]. The existing evidence suggests that these processes are complementary and mutually influential, with inhibiting one pathway potentially alleviating the effects of the other.

In conclusion, our findings illuminate the complex interplay between NGF signaling, inflammation, and lipid peroxidation in the context of preterm birth. Increased NGF levels disrupt homeostasis within the maternal decidua, leading to elevated lipid peroxidation, which collectively contribute to the onset of inflammation and preterm labor. Future investigations may focus on delineating the precise molecular pathways through which NGF and lipid peroxidation interact, paving the way for targeted interventions aimed at improving pregnancy outcomes in at-risk populations.